This one sounds like it comes right out of a science-fiction writer's nightmare. A US intelligence agency wants to develop applications based on the way the human brain makes sense of large amounts of haphazard, partial information.
This week Raytheon BBN Technologies was awarded $3 million by the Intelligence Advanced Research Projects Activity (IARPA) group to explore new methods of modeling what it calls the brain's sensemaking ability. The research could have commercial and military benefits, such as helping the intelligence community analyze fast-moving battlefield video, audio, and text data quickly and accurately, IARPA stated.
More interesting items: The weirdest, wackiest and stupidest sci/tech stories
According to IARPA, which invests in long-range, high-risk/high-payoff research programs, sensemaking refers to the process by which humans are able to generate explanations for data that are otherwise sparse, noisy, and uncertain. It is a core cognitive ability that is central to the work of intelligence analysts, IARPA says. Yet despite its importance, sensemaking remains a poorly understood phenomenon. To date, sensemaking has been studied primarily at the psychological, behavioral and social levels. The models that have emerged from this research, while compelling, remain largely descriptive or qualitative in nature.
IARPA wants more than that. The BBN award is part of the agency's Integrated Cognitive-Neuroscience Architectures for Understanding Sensemaking (ICARUS) program, which by the way is overseen by the Office of the Director of National Intelligence. ICARUS looks to "develop brain-based computational models that explain the fundamental mechanisms of human sensemaking and that demonstrate the strengths and weaknesses of human sensemaking performance."
From IARPA: "By shedding light on the fundamental mechanisms of sensemaking, ICARUS models will enable the Intelligence Community to better predict human-related strengths and failure modes in the intelligence analysis process and will point to new strategies for enhancing analytic tools and methods. Furthermore, ICARUS models may serve to help define a platform for a new generation of automated analysis tools."
ICARUS isn't the only program IARPA is looking into to build computer systems utilizing daily human activities. The agency this month said it was looking to build programs that could aggregate publically available data and develop intelligent systems to predict future events. The program, known as the Open Source Indicators (OSI) will aim to "develop methods for continuous, automated analysis of publicly available data in order to anticipate and/or detect societal disruptions, such as political crises, disease outbreaks, economic instability, resource shortages, and natural disasters," IARPA stated.
In May IARPA said it wanted to build a repository of metaphors. Not just American/English metaphors mind you but those of Iranian Farsi, Mexican Spanish and Russian speakers.
In the end the program should produce a methodology, tools and techniques together with a prototype system that will identify metaphors that provide insight into cultural beliefs. It should also help build structured framework that organizes the metaphors associated with the various dimensions of an analytic problem and build a metaphor repository where all metaphors and related information are captured for future reference and access, IARPA stated.
IARPA also runs the Automated Low-Level Analysis and Description of Diverse Intelligence Video (ALADDIN) program which looks to build and analyze what it calls open source video clips.
Follow Michael Cooney on Twitter: nwwlayer8
Layer 8 Extra
Check out these other hot stories: